Remove dependency on zumba/json-serializer #598
Conversation
Instead of deserializing the data and then comparing it with the actual objects, this serializes the actual object and then compares it with the serialized data. Adds duplicated information and the type of objects are lost, but this allows for the removal of zumba/json-serializer dependency. We can develop a specialized serializer for this or use another method to test this data later on. Signed-off-by: Maurício Meneghini Fauth <mauricio@mfauth.net>
| "KEYWORD_NAME_INDICATORS": [ | ||
| "FROM", | ||
| "SET", | ||
| "WHERE" | ||
| ], | ||
| "OPERATOR_NAME_INDICATORS": [ | ||
| ",", | ||
| "." | ||
| ], |
There was a problem hiding this comment.
can you implement something to drop the constants ?
they add quite some data and will cause all files to have to be updated on a new value added
| "str": "WITH](", | ||
| "len": 6, | ||
| "last": 6, | ||
| "list": { | ||
| "@type": "PhpMyAdmin\\SqlParser\\TokensList", |
There was a problem hiding this comment.
having the type was not too bad, but we can drop it
|
Well, if we can trim down a bit the diff by adding back |
Nothing against this lib. It's because we don't actually need a full serializer for theses tests. I think we don't even need to use json for it. |
|
I just noticed that removing the |
Instead of deserializing the data and then comparing it with the actual objects, this serializes the actual object and then compares it with the serialized data.
Adds duplicated information and the type of objects are lost, but this allows for the removal of zumba/json-serializer dependency.
We can develop a specialized serializer for this or use another method to test this data later on.